Calculation cross entropy for batch of two tensors

Beginners
1

I’d like to calculate cross entropy for batch of two tensors:


x = torch.tensor([[[ 2.1137, -1.3133,  0.7930,  0.3330,  0.9407],
        [-0.8380, -2.0299, -1.1218,  0.3150,  0.4797],
        [-0.7439,  0.0753, -0.1121,  0.0096, -1.2621]]])

y = torch.tensor([[1,2,3]])
    ​
loss = nn.CrossEntropyLoss()(x, y)
    ​

but receive exception:
RuntimeError: Expected target size [1, 5], got [1, 3]
Please explain what is wrong…

2

Try

x = torch.tensor([[ 2.1137, -1.3133, 0.7930, 0.3330, 0.9407],[-0.8380, -2.0299, -1.1218, 0.3150, 0.4797],[-0.7439,0.0753,-0.1121,0.0096,-1.2621]])

y = torch.tensor([1,2,3])
loss = torch.nn.CrossEntropyLoss()(x, y)